A Meta Learning-Based Approach for Zero-Shot Co-Training

The lack of labeled data is one of the main obstacles to the application of machine learning algorithms in a variety of domains. Semi-supervised learning, where additional samples are automatically labeled, is a common and cost-effective approach to address this challenge. A popular semi-supervised...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Guy Zaks, Gilad Katz
Formato:	article
Lenguaje:	EN
Publicado:	IEEE 2021
Materias:	Co-training semi-supervised learning meta-learning Electrical engineering. Electronics. Nuclear engineering TK1-9971
Acceso en línea:	https://doaj.org/article/e8c7cfc74f4940ecba62eff0d32920f3
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:e8c7cfc74f4940ecba62eff0d32920f3
record_format	dspace
spelling	oai:doaj.org-article:e8c7cfc74f4940ecba62eff0d32920f32021-11-09T00:02:52ZA Meta Learning-Based Approach for Zero-Shot Co-Training2169-353610.1109/ACCESS.2021.3116972https://doaj.org/article/e8c7cfc74f4940ecba62eff0d32920f32021-01-01T00:00:00Zhttps://ieeexplore.ieee.org/document/9555627/https://doaj.org/toc/2169-3536The lack of labeled data is one of the main obstacles to the application of machine learning algorithms in a variety of domains. Semi-supervised learning, where additional samples are automatically labeled, is a common and cost-effective approach to address this challenge. A popular semi-supervised labeling approach is co-training, where two views of the data – achieved by the training of two learning models on different feature subsets – iteratively provide each other with additional newly-labeled samples. Despite being effective in many cases, existing co-training algorithms often suffer from low labeling accuracy and a heuristic sample-selection strategy that hurt their performance. We propose <italic>Co</italic>-training using <italic>Met</italic>a-learning (CoMet), a novel approach that addresses many of the shortcomings of existing co-training methods. Instead of employing a greedy labeling approach of individual samples, CoMet evaluates batches of samples and is thus able to select samples that complement each other. Additionally, our approach employs a meta-learning approach that enables it to leverage insights from previously-evaluated datasets and apply these insights to other datasets. Extensive evaluation on 35 datasets shows CoMet significantly outperforms other leading co-training approaches, particularly when the amount of available labeled data is very small. Moreover, our analysis shows that CoMet’s labeling accuracy and consistency of performance are also superior to those of existing approaches.Guy ZaksGilad KatzIEEEarticleCo-trainingsemi-supervised learningmeta-learningElectrical engineering. Electronics. Nuclear engineeringTK1-9971ENIEEE Access, Vol 9, Pp 146653-146666 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Co-training semi-supervised learning meta-learning Electrical engineering. Electronics. Nuclear engineering TK1-9971
spellingShingle	Co-training semi-supervised learning meta-learning Electrical engineering. Electronics. Nuclear engineering TK1-9971 Guy Zaks Gilad Katz A Meta Learning-Based Approach for Zero-Shot Co-Training
description	The lack of labeled data is one of the main obstacles to the application of machine learning algorithms in a variety of domains. Semi-supervised learning, where additional samples are automatically labeled, is a common and cost-effective approach to address this challenge. A popular semi-supervised labeling approach is co-training, where two views of the data – achieved by the training of two learning models on different feature subsets – iteratively provide each other with additional newly-labeled samples. Despite being effective in many cases, existing co-training algorithms often suffer from low labeling accuracy and a heuristic sample-selection strategy that hurt their performance. We propose <italic>Co</italic>-training using <italic>Met</italic>a-learning (CoMet), a novel approach that addresses many of the shortcomings of existing co-training methods. Instead of employing a greedy labeling approach of individual samples, CoMet evaluates batches of samples and is thus able to select samples that complement each other. Additionally, our approach employs a meta-learning approach that enables it to leverage insights from previously-evaluated datasets and apply these insights to other datasets. Extensive evaluation on 35 datasets shows CoMet significantly outperforms other leading co-training approaches, particularly when the amount of available labeled data is very small. Moreover, our analysis shows that CoMet’s labeling accuracy and consistency of performance are also superior to those of existing approaches.
format	article
author	Guy Zaks Gilad Katz
author_facet	Guy Zaks Gilad Katz
author_sort	Guy Zaks
title	A Meta Learning-Based Approach for Zero-Shot Co-Training
title_short	A Meta Learning-Based Approach for Zero-Shot Co-Training
title_full	A Meta Learning-Based Approach for Zero-Shot Co-Training
title_fullStr	A Meta Learning-Based Approach for Zero-Shot Co-Training
title_full_unstemmed	A Meta Learning-Based Approach for Zero-Shot Co-Training
title_sort	meta learning-based approach for zero-shot co-training
publisher	IEEE
publishDate	2021
url	https://doaj.org/article/e8c7cfc74f4940ecba62eff0d32920f3
work_keys_str_mv	AT guyzaks ametalearningbasedapproachforzeroshotcotraining AT giladkatz ametalearningbasedapproachforzeroshotcotraining AT guyzaks metalearningbasedapproachforzeroshotcotraining AT giladkatz metalearningbasedapproachforzeroshotcotraining
_version_	1718441434268303360

A Meta Learning-Based Approach for Zero-Shot Co-Training

Ejemplares similares